random.c 47.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3
/*
 * random.c -- A strong random number generator
 *
4
 * Copyright Matt Mackall <mpm@selenic.com>, 2003, 2004, 2005
L
Linus Torvalds 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
 *
 * Copyright Theodore Ts'o, 1994, 1995, 1996, 1997, 1998, 1999.  All
 * rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, and the entire permission notice in its entirety,
 *    including the disclaimer of warranties.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. The name of the author may not be used to endorse or promote
 *    products derived from this software without specific prior
 *    written permission.
 *
 * ALTERNATIVELY, this product may be distributed under the terms of
 * the GNU General Public License, in which case the provisions of the GPL are
 * required INSTEAD OF the above restrictions.  (This clause is
 * necessary due to a potential bad interaction between the GPL and
 * the restrictions contained in a BSD-style copyright.)
 *
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, ALL OF
 * WHICH ARE HEREBY DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
 * OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
 * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
 * USE OF THIS SOFTWARE, EVEN IF NOT ADVISED OF THE POSSIBILITY OF SUCH
 * DAMAGE.
 */

/*
 * (now, with legal B.S. out of the way.....)
 *
 * This routine gathers environmental noise from device drivers, etc.,
 * and returns good random numbers, suitable for cryptographic use.
 * Besides the obvious cryptographic uses, these numbers are also good
 * for seeding TCP sequence numbers, and other places where it is
 * desirable to have numbers which are not only random, but hard to
 * predict by an attacker.
 *
 * Theory of operation
 * ===================
 *
 * Computers are very predictable devices.  Hence it is extremely hard
 * to produce truly random numbers on a computer --- as opposed to
 * pseudo-random numbers, which can easily generated by using a
 * algorithm.  Unfortunately, it is very easy for attackers to guess
 * the sequence of pseudo-random number generators, and for some
 * applications this is not acceptable.  So instead, we must try to
 * gather "environmental noise" from the computer's environment, which
 * must be hard for outside attackers to observe, and use that to
 * generate random numbers.  In a Unix environment, this is best done
 * from inside the kernel.
 *
 * Sources of randomness from the environment include inter-keyboard
 * timings, inter-interrupt timings from some interrupts, and other
 * events which are both (a) non-deterministic and (b) hard for an
 * outside observer to measure.  Randomness from these sources are
 * added to an "entropy pool", which is mixed using a CRC-like function.
 * This is not cryptographically strong, but it is adequate assuming
 * the randomness is not chosen maliciously, and it is fast enough that
 * the overhead of doing it on every interrupt is very reasonable.
 * As random bytes are mixed into the entropy pool, the routines keep
 * an *estimate* of how many bits of randomness have been stored into
 * the random number generator's internal state.
 *
 * When random bytes are desired, they are obtained by taking the SHA
 * hash of the contents of the "entropy pool".  The SHA hash avoids
 * exposing the internal state of the entropy pool.  It is believed to
 * be computationally infeasible to derive any useful information
 * about the input of SHA from its output.  Even if it is possible to
 * analyze SHA in some clever way, as long as the amount of data
 * returned from the generator is less than the inherent entropy in
 * the pool, the output data is totally unpredictable.  For this
 * reason, the routine decreases its internal estimate of how many
 * bits of "true randomness" are contained in the entropy pool as it
 * outputs random numbers.
 *
 * If this estimate goes to zero, the routine can still generate
 * random numbers; however, an attacker may (at least in theory) be
 * able to infer the future output of the generator from prior
 * outputs.  This requires successful cryptanalysis of SHA, which is
 * not believed to be feasible, but there is a remote possibility.
 * Nonetheless, these numbers should be useful for the vast majority
 * of purposes.
 *
 * Exported interfaces ---- output
 * ===============================
 *
 * There are three exported interfaces; the first is one designed to
 * be used from within the kernel:
 *
 * 	void get_random_bytes(void *buf, int nbytes);
 *
 * This interface will return the requested number of random bytes,
 * and place it in the requested buffer.
 *
 * The two other interfaces are two character devices /dev/random and
 * /dev/urandom.  /dev/random is suitable for use when very high
 * quality randomness is desired (for example, for key generation or
 * one-time pads), as it will only return a maximum of the number of
 * bits of randomness (as estimated by the random number generator)
 * contained in the entropy pool.
 *
 * The /dev/urandom device does not have this limit, and will return
 * as many bytes as are requested.  As more and more random bytes are
 * requested without giving time for the entropy pool to recharge,
 * this will result in random numbers that are merely cryptographically
 * strong.  For many applications, however, this is acceptable.
 *
 * Exported interfaces ---- input
 * ==============================
 *
 * The current exported interfaces for gathering environmental noise
 * from the devices are:
 *
128
 *	void add_device_randomness(const void *buf, unsigned int size);
L
Linus Torvalds 已提交
129 130
 * 	void add_input_randomness(unsigned int type, unsigned int code,
 *                                unsigned int value);
131
 *	void add_interrupt_randomness(int irq, int irq_flags);
132
 * 	void add_disk_randomness(struct gendisk *disk);
L
Linus Torvalds 已提交
133
 *
134 135 136 137 138 139 140 141
 * add_device_randomness() is for adding data to the random pool that
 * is likely to differ between two devices (or possibly even per boot).
 * This would be things like MAC addresses or serial numbers, or the
 * read-out of the RTC. This does *not* add any actual entropy to the
 * pool, but it initializes the pool to different values for devices
 * that might otherwise be identical and have very little entropy
 * available to them (particularly common in the embedded world).
 *
L
Linus Torvalds 已提交
142 143 144
 * add_input_randomness() uses the input layer interrupt timing, as well as
 * the event type information from the hardware.
 *
145 146 147
 * add_interrupt_randomness() uses the interrupt timing as random
 * inputs to the entropy pool. Using the cycle counters and the irq source
 * as inputs, it feeds the randomness roughly once a second.
148 149 150 151 152 153
 *
 * add_disk_randomness() uses what amounts to the seek time of block
 * layer request events, on a per-disk_devt basis, as input to the
 * entropy pool. Note that high-speed solid state drives with very low
 * seek times do not make for good sources of entropy, as their seek
 * times are usually fairly consistent.
L
Linus Torvalds 已提交
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250
 *
 * All of these routines try to estimate how many bits of randomness a
 * particular randomness source.  They do this by keeping track of the
 * first and second order deltas of the event timings.
 *
 * Ensuring unpredictability at system startup
 * ============================================
 *
 * When any operating system starts up, it will go through a sequence
 * of actions that are fairly predictable by an adversary, especially
 * if the start-up does not involve interaction with a human operator.
 * This reduces the actual number of bits of unpredictability in the
 * entropy pool below the value in entropy_count.  In order to
 * counteract this effect, it helps to carry information in the
 * entropy pool across shut-downs and start-ups.  To do this, put the
 * following lines an appropriate script which is run during the boot
 * sequence:
 *
 *	echo "Initializing random number generator..."
 *	random_seed=/var/run/random-seed
 *	# Carry a random seed from start-up to start-up
 *	# Load and then save the whole entropy pool
 *	if [ -f $random_seed ]; then
 *		cat $random_seed >/dev/urandom
 *	else
 *		touch $random_seed
 *	fi
 *	chmod 600 $random_seed
 *	dd if=/dev/urandom of=$random_seed count=1 bs=512
 *
 * and the following lines in an appropriate script which is run as
 * the system is shutdown:
 *
 *	# Carry a random seed from shut-down to start-up
 *	# Save the whole entropy pool
 *	echo "Saving random seed..."
 *	random_seed=/var/run/random-seed
 *	touch $random_seed
 *	chmod 600 $random_seed
 *	dd if=/dev/urandom of=$random_seed count=1 bs=512
 *
 * For example, on most modern systems using the System V init
 * scripts, such code fragments would be found in
 * /etc/rc.d/init.d/random.  On older Linux systems, the correct script
 * location might be in /etc/rcb.d/rc.local or /etc/rc.d/rc.0.
 *
 * Effectively, these commands cause the contents of the entropy pool
 * to be saved at shut-down time and reloaded into the entropy pool at
 * start-up.  (The 'dd' in the addition to the bootup script is to
 * make sure that /etc/random-seed is different for every start-up,
 * even if the system crashes without executing rc.0.)  Even with
 * complete knowledge of the start-up activities, predicting the state
 * of the entropy pool requires knowledge of the previous history of
 * the system.
 *
 * Configuring the /dev/random driver under Linux
 * ==============================================
 *
 * The /dev/random driver under Linux uses minor numbers 8 and 9 of
 * the /dev/mem major number (#1).  So if your system does not have
 * /dev/random and /dev/urandom created already, they can be created
 * by using the commands:
 *
 * 	mknod /dev/random c 1 8
 * 	mknod /dev/urandom c 1 9
 *
 * Acknowledgements:
 * =================
 *
 * Ideas for constructing this random number generator were derived
 * from Pretty Good Privacy's random number generator, and from private
 * discussions with Phil Karn.  Colin Plumb provided a faster random
 * number generator, which speed up the mixing function of the entropy
 * pool, taken from PGPfone.  Dale Worley has also contributed many
 * useful ideas and suggestions to improve this driver.
 *
 * Any flaws in the design are solely my responsibility, and should
 * not be attributed to the Phil, Colin, or any of authors of PGP.
 *
 * Further background information on this topic may be obtained from
 * RFC 1750, "Randomness Recommendations for Security", by Donald
 * Eastlake, Steve Crocker, and Jeff Schiller.
 */

#include <linux/utsname.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/major.h>
#include <linux/string.h>
#include <linux/fcntl.h>
#include <linux/slab.h>
#include <linux/random.h>
#include <linux/poll.h>
#include <linux/init.h>
#include <linux/fs.h>
#include <linux/genhd.h>
#include <linux/interrupt.h>
251
#include <linux/mm.h>
L
Linus Torvalds 已提交
252 253 254
#include <linux/spinlock.h>
#include <linux/percpu.h>
#include <linux/cryptohash.h>
255
#include <linux/fips.h>
256
#include <linux/ptrace.h>
257
#include <linux/kmemcheck.h>
L
Linus Torvalds 已提交
258

259 260 261 262
#ifdef CONFIG_GENERIC_HARDIRQS
# include <linux/irq.h>
#endif

L
Linus Torvalds 已提交
263 264 265
#include <asm/processor.h>
#include <asm/uaccess.h>
#include <asm/irq.h>
266
#include <asm/irq_regs.h>
L
Linus Torvalds 已提交
267 268
#include <asm/io.h>

269 270 271
#define CREATE_TRACE_POINTS
#include <trace/events/random.h>

L
Linus Torvalds 已提交
272 273 274
/*
 * Configuration information
 */
275 276 277 278 279 280
#define INPUT_POOL_SHIFT	12
#define INPUT_POOL_WORDS	(1 << (INPUT_POOL_SHIFT-5))
#define OUTPUT_POOL_SHIFT	10
#define OUTPUT_POOL_WORDS	(1 << (OUTPUT_POOL_SHIFT-5))
#define SEC_XFER_SIZE		512
#define EXTRACT_SIZE		10
L
Linus Torvalds 已提交
281

282 283
#define LONGS(x) (((x) + sizeof(unsigned long) - 1)/sizeof(unsigned long))

284 285 286 287 288
/*
 * To allow fractional bits to be tracked, the following fields contain
 * this many fractional bits:
 *
 * entropy_count, trickle_thresh
289 290 291
 *
 * 2*(ENTROPY_SHIFT + log2(poolbits)) must <= 31, or the multiply in
 * credit_entropy_bits() needs to be 64 bits wide.
292 293 294 295
 */
#define ENTROPY_SHIFT 3
#define ENTROPY_BITS(r) ((r)->entropy_count >> ENTROPY_SHIFT)

L
Linus Torvalds 已提交
296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
/*
 * The minimum number of bits of entropy before we wake up a read on
 * /dev/random.  Should be enough to do a significant reseed.
 */
static int random_read_wakeup_thresh = 64;

/*
 * If the entropy count falls under this number of bits, then we
 * should wake up processes which are selecting or polling on write
 * access to /dev/random.
 */
static int random_write_wakeup_thresh = 128;

/*
 * When the input pool goes over trickle_thresh, start dropping most
 * samples to avoid wasting CPU time and reduce lock contention.
 */
313
static const int trickle_thresh = (INPUT_POOL_WORDS * 28) << ENTROPY_SHIFT;
L
Linus Torvalds 已提交
314

315
static DEFINE_PER_CPU(int, trickle_count);
L
Linus Torvalds 已提交
316 317 318 319 320 321 322 323 324

/*
 * A pool of size .poolwords is stirred with a primitive polynomial
 * of degree .poolwords over GF(2).  The taps for various sizes are
 * defined below.  They are chosen to be evenly spaced (minimum RMS
 * distance from evenly spaced; the numbers in the comments are a
 * scaled squared error sum) except for the last tap, which is 1 to
 * get the twisting happening as fast as possible.
 */
325

L
Linus Torvalds 已提交
326
static struct poolinfo {
327 328
	int poolbitshift, poolwords, poolbytes, poolbits, poolfracbits;
#define S(x) ilog2(x)+5, (x), (x)*4, (x)*32, (x) << (ENTROPY_SHIFT+5)
L
Linus Torvalds 已提交
329 330 331
	int tap1, tap2, tap3, tap4, tap5;
} poolinfo_table[] = {
	/* x^128 + x^103 + x^76 + x^51 +x^25 + x + 1 -- 105 */
332
	{ S(128),	103,	76,	51,	25,	1 },
L
Linus Torvalds 已提交
333
	/* x^32 + x^26 + x^20 + x^14 + x^7 + x + 1 -- 15 */
334
	{ S(32),	26,	20,	14,	7,	1 },
L
Linus Torvalds 已提交
335 336
#if 0
	/* x^2048 + x^1638 + x^1231 + x^819 + x^411 + x + 1  -- 115 */
337
	{ S(2048),	1638,	1231,	819,	411,	1 },
L
Linus Torvalds 已提交
338 339

	/* x^1024 + x^817 + x^615 + x^412 + x^204 + x + 1 -- 290 */
340
	{ S(1024),	817,	615,	412,	204,	1 },
L
Linus Torvalds 已提交
341 342

	/* x^1024 + x^819 + x^616 + x^410 + x^207 + x^2 + 1 -- 115 */
343
	{ S(1024),	819,	616,	410,	207,	2 },
L
Linus Torvalds 已提交
344 345

	/* x^512 + x^411 + x^308 + x^208 + x^104 + x + 1 -- 225 */
346
	{ S(512),	411,	308,	208,	104,	1 },
L
Linus Torvalds 已提交
347 348

	/* x^512 + x^409 + x^307 + x^206 + x^102 + x^2 + 1 -- 95 */
349
	{ S(512),	409,	307,	206,	102,	2 },
L
Linus Torvalds 已提交
350
	/* x^512 + x^409 + x^309 + x^205 + x^103 + x^2 + 1 -- 95 */
351
	{ S(512),	409,	309,	205,	103,	2 },
L
Linus Torvalds 已提交
352 353

	/* x^256 + x^205 + x^155 + x^101 + x^52 + x + 1 -- 125 */
354
	{ S(256),	205,	155,	101,	52,	1 },
L
Linus Torvalds 已提交
355 356

	/* x^128 + x^103 + x^78 + x^51 + x^27 + x^2 + 1 -- 70 */
357
	{ S(128),	103,	78,	51,	27,	2 },
L
Linus Torvalds 已提交
358 359

	/* x^64 + x^52 + x^39 + x^26 + x^14 + x + 1 -- 15 */
360
	{ S(64),	52,	39,	26,	14,	1 },
L
Linus Torvalds 已提交
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
#endif
};

/*
 * For the purposes of better mixing, we use the CRC-32 polynomial as
 * well to make a twisted Generalized Feedback Shift Reigster
 *
 * (See M. Matsumoto & Y. Kurita, 1992.  Twisted GFSR generators.  ACM
 * Transactions on Modeling and Computer Simulation 2(3):179-194.
 * Also see M. Matsumoto & Y. Kurita, 1994.  Twisted GFSR generators
 * II.  ACM Transactions on Mdeling and Computer Simulation 4:254-266)
 *
 * Thanks to Colin Plumb for suggesting this.
 *
 * We have not analyzed the resultant polynomial to prove it primitive;
 * in fact it almost certainly isn't.  Nonetheless, the irreducible factors
 * of a random large-degree polynomial over GF(2) are more than large enough
 * that periodicity is not a concern.
 *
 * The input hash is much less sensitive than the output hash.  All
 * that we want of it is that it be a good non-cryptographic hash;
 * i.e. it not produce collisions when fed "random" data of the sort
 * we expect to see.  As long as the pool state differs for different
 * inputs, we have preserved the input entropy and done a good job.
 * The fact that an intelligent attacker can construct inputs that
 * will produce controlled alterations to the pool's state is not
 * important because we don't consider such inputs to contribute any
 * randomness.  The only property we need with respect to them is that
 * the attacker can't increase his/her knowledge of the pool's state.
 * Since all additions are reversible (knowing the final state and the
 * input, you can reconstruct the initial state), if an attacker has
 * any uncertainty about the initial state, he/she can only shuffle
 * that uncertainty about, but never cause any collisions (which would
 * decrease the uncertainty).
 *
 * The chosen system lets the state of the pool be (essentially) the input
 * modulo the generator polymnomial.  Now, for random primitive polynomials,
 * this is a universal class of hash functions, meaning that the chance
 * of a collision is limited by the attacker's knowledge of the generator
 * polynomail, so if it is chosen at random, an attacker can never force
 * a collision.  Here, we use a fixed polynomial, but we *can* assume that
 * ###--> it is unknown to the processes generating the input entropy. <-###
 * Because of this important property, this is a good, collision-resistant
 * hash; hash collisions will occur no more often than chance.
 */

/*
 * Static global variables
 */
static DECLARE_WAIT_QUEUE_HEAD(random_read_wait);
static DECLARE_WAIT_QUEUE_HEAD(random_write_wait);
412
static struct fasync_struct *fasync;
L
Linus Torvalds 已提交
413

414
static bool debug;
L
Linus Torvalds 已提交
415
module_param(debug, bool, 0644);
416 417 418 419 420 421 422 423
#define DEBUG_ENT(fmt, arg...) do { \
	if (debug) \
		printk(KERN_DEBUG "random %04d %04d %04d: " \
		fmt,\
		input_pool.entropy_count,\
		blocking_pool.entropy_count,\
		nonblocking_pool.entropy_count,\
		## arg); } while (0)
L
Linus Torvalds 已提交
424 425 426 427 428 429 430 431 432 433

/**********************************************************************
 *
 * OS independent entropy store.   Here are the functions which handle
 * storing entropy in an entropy pool.
 *
 **********************************************************************/

struct entropy_store;
struct entropy_store {
434
	/* read-only data: */
435
	const struct poolinfo *poolinfo;
L
Linus Torvalds 已提交
436 437 438
	__u32 *pool;
	const char *name;
	struct entropy_store *pull;
439
	int limit;
L
Linus Torvalds 已提交
440 441

	/* read-write data: */
442
	spinlock_t lock;
L
Linus Torvalds 已提交
443
	unsigned add_ptr;
444
	unsigned input_rotate;
445
	int entropy_count;
446 447
	int entropy_total;
	unsigned int initialized:1;
448
	bool last_data_init;
M
Matt Mackall 已提交
449
	__u8 last_data[EXTRACT_SIZE];
L
Linus Torvalds 已提交
450 451 452 453 454 455 456 457 458 459
};

static __u32 input_pool_data[INPUT_POOL_WORDS];
static __u32 blocking_pool_data[OUTPUT_POOL_WORDS];
static __u32 nonblocking_pool_data[OUTPUT_POOL_WORDS];

static struct entropy_store input_pool = {
	.poolinfo = &poolinfo_table[0],
	.name = "input",
	.limit = 1,
460
	.lock = __SPIN_LOCK_UNLOCKED(input_pool.lock),
L
Linus Torvalds 已提交
461 462 463 464 465 466 467 468
	.pool = input_pool_data
};

static struct entropy_store blocking_pool = {
	.poolinfo = &poolinfo_table[1],
	.name = "blocking",
	.limit = 1,
	.pull = &input_pool,
469
	.lock = __SPIN_LOCK_UNLOCKED(blocking_pool.lock),
L
Linus Torvalds 已提交
470 471 472 473 474 475 476
	.pool = blocking_pool_data
};

static struct entropy_store nonblocking_pool = {
	.poolinfo = &poolinfo_table[1],
	.name = "nonblocking",
	.pull = &input_pool,
477
	.lock = __SPIN_LOCK_UNLOCKED(nonblocking_pool.lock),
L
Linus Torvalds 已提交
478 479 480
	.pool = nonblocking_pool_data
};

481 482 483 484
static __u32 const twist_table[8] = {
	0x00000000, 0x3b6e20c8, 0x76dc4190, 0x4db26158,
	0xedb88320, 0xd6d6a3e8, 0x9b64c2b0, 0xa00ae278 };

L
Linus Torvalds 已提交
485
/*
486
 * This function adds bytes into the entropy "pool".  It does not
L
Linus Torvalds 已提交
487
 * update the entropy estimate.  The caller should call
488
 * credit_entropy_bits if this is appropriate.
L
Linus Torvalds 已提交
489 490 491 492 493 494
 *
 * The pool is stirred with a primitive polynomial of the appropriate
 * degree, and then twisted.  We twist by three bits at a time because
 * it's cheap to do so and helps slightly in the expected case where
 * the entropy is concentrated in the low-order bits.
 */
495 496
static void _mix_pool_bytes(struct entropy_store *r, const void *in,
			    int nbytes, __u8 out[64])
L
Linus Torvalds 已提交
497
{
M
Matt Mackall 已提交
498
	unsigned long i, j, tap1, tap2, tap3, tap4, tap5;
499
	int input_rotate;
L
Linus Torvalds 已提交
500
	int wordmask = r->poolinfo->poolwords - 1;
501
	const char *bytes = in;
502
	__u32 w;
L
Linus Torvalds 已提交
503 504 505 506 507 508 509

	tap1 = r->poolinfo->tap1;
	tap2 = r->poolinfo->tap2;
	tap3 = r->poolinfo->tap3;
	tap4 = r->poolinfo->tap4;
	tap5 = r->poolinfo->tap5;

510 511 512
	smp_rmb();
	input_rotate = ACCESS_ONCE(r->input_rotate);
	i = ACCESS_ONCE(r->add_ptr);
L
Linus Torvalds 已提交
513

514 515 516
	/* mix one byte at a time to simplify size handling and churn faster */
	while (nbytes--) {
		w = rol32(*bytes++, input_rotate & 31);
M
Matt Mackall 已提交
517
		i = (i - 1) & wordmask;
L
Linus Torvalds 已提交
518 519

		/* XOR in the various taps */
M
Matt Mackall 已提交
520
		w ^= r->pool[i];
L
Linus Torvalds 已提交
521 522 523 524 525
		w ^= r->pool[(i + tap1) & wordmask];
		w ^= r->pool[(i + tap2) & wordmask];
		w ^= r->pool[(i + tap3) & wordmask];
		w ^= r->pool[(i + tap4) & wordmask];
		w ^= r->pool[(i + tap5) & wordmask];
M
Matt Mackall 已提交
526 527

		/* Mix the result back in with a twist */
L
Linus Torvalds 已提交
528
		r->pool[i] = (w >> 3) ^ twist_table[w & 7];
529 530 531 532 533 534 535 536

		/*
		 * Normally, we add 7 bits of rotation to the pool.
		 * At the beginning of the pool, add an extra 7 bits
		 * rotation, so that successive passes spread the
		 * input bits across the pool evenly.
		 */
		input_rotate += i ? 7 : 14;
L
Linus Torvalds 已提交
537 538
	}

539 540 541
	ACCESS_ONCE(r->input_rotate) = input_rotate;
	ACCESS_ONCE(r->add_ptr) = i;
	smp_wmb();
L
Linus Torvalds 已提交
542

M
Matt Mackall 已提交
543 544
	if (out)
		for (j = 0; j < 16; j++)
545
			((__u32 *)out)[j] = r->pool[(i - j) & wordmask];
L
Linus Torvalds 已提交
546 547
}

548
static void __mix_pool_bytes(struct entropy_store *r, const void *in,
549
			     int nbytes, __u8 out[64])
550 551 552 553 554 555 556
{
	trace_mix_pool_bytes_nolock(r->name, nbytes, _RET_IP_);
	_mix_pool_bytes(r, in, nbytes, out);
}

static void mix_pool_bytes(struct entropy_store *r, const void *in,
			   int nbytes, __u8 out[64])
L
Linus Torvalds 已提交
557
{
558 559
	unsigned long flags;

560
	trace_mix_pool_bytes(r->name, nbytes, _RET_IP_);
561
	spin_lock_irqsave(&r->lock, flags);
562
	_mix_pool_bytes(r, in, nbytes, out);
563
	spin_unlock_irqrestore(&r->lock, flags);
L
Linus Torvalds 已提交
564 565
}

566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
struct fast_pool {
	__u32		pool[4];
	unsigned long	last;
	unsigned short	count;
	unsigned char	rotate;
	unsigned char	last_timer_intr;
};

/*
 * This is a fast mixing routine used by the interrupt randomness
 * collector.  It's hardcoded for an 128 bit pool and assumes that any
 * locks that might be needed are taken by the caller.
 */
static void fast_mix(struct fast_pool *f, const void *in, int nbytes)
{
	const char	*bytes = in;
	__u32		w;
	unsigned	i = f->count;
	unsigned	input_rotate = f->rotate;

	while (nbytes--) {
		w = rol32(*bytes++, input_rotate & 31) ^ f->pool[i & 3] ^
			f->pool[(i + 1) & 3];
		f->pool[i & 3] = (w >> 3) ^ twist_table[w & 7];
		input_rotate += (i++ & 3) ? 7 : 14;
	}
	f->count = i;
	f->rotate = input_rotate;
}

L
Linus Torvalds 已提交
596
/*
597 598 599
 * Credit (or debit) the entropy store with n bits of entropy.
 * Use credit_entropy_bits_safe() if the value comes from userspace
 * or otherwise should be checked for extreme values.
L
Linus Torvalds 已提交
600
 */
601
static void credit_entropy_bits(struct entropy_store *r, int nbits)
L
Linus Torvalds 已提交
602
{
603
	int entropy_count, orig;
604 605
	const int pool_size = r->poolinfo->poolfracbits;
	int nfrac = nbits << ENTROPY_SHIFT;
L
Linus Torvalds 已提交
606

607 608 609 610
	if (!nbits)
		return;

	DEBUG_ENT("added %d entropy credits to %s\n", nbits, r->name);
611 612
retry:
	entropy_count = orig = ACCESS_ONCE(r->entropy_count);
613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650
	if (nfrac < 0) {
		/* Debit */
		entropy_count += nfrac;
	} else {
		/*
		 * Credit: we have to account for the possibility of
		 * overwriting already present entropy.	 Even in the
		 * ideal case of pure Shannon entropy, new contributions
		 * approach the full value asymptotically:
		 *
		 * entropy <- entropy + (pool_size - entropy) *
		 *	(1 - exp(-add_entropy/pool_size))
		 *
		 * For add_entropy <= pool_size/2 then
		 * (1 - exp(-add_entropy/pool_size)) >=
		 *    (add_entropy/pool_size)*0.7869...
		 * so we can approximate the exponential with
		 * 3/4*add_entropy/pool_size and still be on the
		 * safe side by adding at most pool_size/2 at a time.
		 *
		 * The use of pool_size-2 in the while statement is to
		 * prevent rounding artifacts from making the loop
		 * arbitrarily long; this limits the loop to log2(pool_size)*2
		 * turns no matter how large nbits is.
		 */
		int pnfrac = nfrac;
		const int s = r->poolinfo->poolbitshift + ENTROPY_SHIFT + 2;
		/* The +2 corresponds to the /4 in the denominator */

		do {
			unsigned int anfrac = min(pnfrac, pool_size/2);
			unsigned int add =
				((pool_size - entropy_count)*anfrac*3) >> s;

			entropy_count += add;
			pnfrac -= anfrac;
		} while (unlikely(entropy_count < pool_size-2 && pnfrac));
	}
651

652
	if (entropy_count < 0) {
653
		DEBUG_ENT("negative entropy/overflow\n");
654
		entropy_count = 0;
655 656
	} else if (entropy_count > pool_size)
		entropy_count = pool_size;
657 658
	if (cmpxchg(&r->entropy_count, orig, entropy_count) != orig)
		goto retry;
L
Linus Torvalds 已提交
659

660 661 662 663 664 665
	if (!r->initialized && nbits > 0) {
		r->entropy_total += nbits;
		if (r->entropy_total > 128)
			r->initialized = 1;
	}

666 667
	trace_credit_entropy_bits(r->name, nbits,
				  entropy_count >> ENTROPY_SHIFT,
668 669
				  r->entropy_total, _RET_IP_);

M
Matt Mackall 已提交
670
	/* should we wake readers? */
671 672
	if (r == &input_pool &&
	    (entropy_count >> ENTROPY_SHIFT) >= random_read_wakeup_thresh) {
M
Matt Mackall 已提交
673
		wake_up_interruptible(&random_read_wait);
674 675
		kill_fasync(&fasync, SIGIO, POLL_IN);
	}
L
Linus Torvalds 已提交
676 677
}

678 679 680 681 682 683 684 685 686 687 688
static void credit_entropy_bits_safe(struct entropy_store *r, int nbits)
{
	const int nbits_max = (int)(~0U >> (ENTROPY_SHIFT + 1));

	/* Cap the value to avoid overflows */
	nbits = min(nbits,  nbits_max);
	nbits = max(nbits, -nbits_max);

	credit_entropy_bits(r, nbits);
}

L
Linus Torvalds 已提交
689 690 691 692 693 694 695 696 697
/*********************************************************************
 *
 * Entropy input management
 *
 *********************************************************************/

/* There is one of these per entropy source */
struct timer_rand_state {
	cycles_t last_time;
698
	long last_delta, last_delta2;
L
Linus Torvalds 已提交
699 700 701
	unsigned dont_count_entropy:1;
};

702 703 704 705 706 707 708 709 710 711
/*
 * Add device- or boot-specific data to the input and nonblocking
 * pools to help initialize them to unique values.
 *
 * None of this adds any entropy, it is meant to avoid the
 * problem of the nonblocking pool having similar initial state
 * across largely identical devices.
 */
void add_device_randomness(const void *buf, unsigned int size)
{
712
	unsigned long time = random_get_entropy() ^ jiffies;
713 714 715 716 717 718 719 720

	mix_pool_bytes(&input_pool, buf, size, NULL);
	mix_pool_bytes(&input_pool, &time, sizeof(time), NULL);
	mix_pool_bytes(&nonblocking_pool, buf, size, NULL);
	mix_pool_bytes(&nonblocking_pool, &time, sizeof(time), NULL);
}
EXPORT_SYMBOL(add_device_randomness);

721 722
static struct timer_rand_state input_timer_state;

L
Linus Torvalds 已提交
723 724 725 726 727 728 729 730 731 732 733 734 735 736
/*
 * This function adds entropy to the entropy "pool" by using timing
 * delays.  It uses the timer_rand_state structure to make an estimate
 * of how many bits of entropy this call has added to the pool.
 *
 * The number "num" is also added to the pool - it should somehow describe
 * the type of event which just happened.  This is currently 0-255 for
 * keyboard scan codes, and 256 upwards for interrupts.
 *
 */
static void add_timer_randomness(struct timer_rand_state *state, unsigned num)
{
	struct {
		long jiffies;
737
		unsigned cycles;
L
Linus Torvalds 已提交
738 739 740 741 742 743
		unsigned num;
	} sample;
	long delta, delta2, delta3;

	preempt_disable();
	/* if over the trickle threshold, use only 1 in 4096 samples */
744
	if (ENTROPY_BITS(&input_pool) > trickle_thresh &&
745
	    ((__this_cpu_inc_return(trickle_count) - 1) & 0xfff))
L
Linus Torvalds 已提交
746 747 748
		goto out;

	sample.jiffies = jiffies;
749
	sample.cycles = random_get_entropy();
L
Linus Torvalds 已提交
750
	sample.num = num;
751
	mix_pool_bytes(&input_pool, &sample, sizeof(sample), NULL);
L
Linus Torvalds 已提交
752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784

	/*
	 * Calculate number of bits of randomness we probably added.
	 * We take into account the first, second and third-order deltas
	 * in order to make our estimate.
	 */

	if (!state->dont_count_entropy) {
		delta = sample.jiffies - state->last_time;
		state->last_time = sample.jiffies;

		delta2 = delta - state->last_delta;
		state->last_delta = delta;

		delta3 = delta2 - state->last_delta2;
		state->last_delta2 = delta2;

		if (delta < 0)
			delta = -delta;
		if (delta2 < 0)
			delta2 = -delta2;
		if (delta3 < 0)
			delta3 = -delta3;
		if (delta > delta2)
			delta = delta2;
		if (delta > delta3)
			delta = delta3;

		/*
		 * delta is now minimum absolute delta.
		 * Round down by 1 bit on general principles,
		 * and limit entropy entimate to 12 bits.
		 */
785 786
		credit_entropy_bits(&input_pool,
				    min_t(int, fls(delta>>1), 11));
L
Linus Torvalds 已提交
787 788 789 790 791
	}
out:
	preempt_enable();
}

792
void add_input_randomness(unsigned int type, unsigned int code,
L
Linus Torvalds 已提交
793 794 795 796 797 798 799 800 801 802 803 804 805
				 unsigned int value)
{
	static unsigned char last_value;

	/* ignore autorepeat and the like */
	if (value == last_value)
		return;

	DEBUG_ENT("input event\n");
	last_value = value;
	add_timer_randomness(&input_timer_state,
			     (type << 4) ^ code ^ (code >> 4) ^ value);
}
806
EXPORT_SYMBOL_GPL(add_input_randomness);
L
Linus Torvalds 已提交
807

808 809 810
static DEFINE_PER_CPU(struct fast_pool, irq_randomness);

void add_interrupt_randomness(int irq, int irq_flags)
L
Linus Torvalds 已提交
811
{
812 813 814 815
	struct entropy_store	*r;
	struct fast_pool	*fast_pool = &__get_cpu_var(irq_randomness);
	struct pt_regs		*regs = get_irq_regs();
	unsigned long		now = jiffies;
816
	__u32			input[4], cycles = random_get_entropy();
817 818 819 820 821 822 823 824

	input[0] = cycles ^ jiffies;
	input[1] = irq;
	if (regs) {
		__u64 ip = instruction_pointer(regs);
		input[2] = ip;
		input[3] = ip >> 32;
	}
825

826
	fast_mix(fast_pool, input, sizeof(input));
827

828 829
	if ((fast_pool->count & 1023) &&
	    !time_after(now, fast_pool->last + HZ))
L
Linus Torvalds 已提交
830 831
		return;

832 833 834
	fast_pool->last = now;

	r = nonblocking_pool.initialized ? &input_pool : &nonblocking_pool;
835
	__mix_pool_bytes(r, &fast_pool->pool, sizeof(fast_pool->pool), NULL);
836 837 838 839 840 841 842 843 844 845 846 847 848 849
	/*
	 * If we don't have a valid cycle counter, and we see
	 * back-to-back timer interrupts, then skip giving credit for
	 * any entropy.
	 */
	if (cycles == 0) {
		if (irq_flags & __IRQF_TIMER) {
			if (fast_pool->last_timer_intr)
				return;
			fast_pool->last_timer_intr = 1;
		} else
			fast_pool->last_timer_intr = 0;
	}
	credit_entropy_bits(r, 1);
L
Linus Torvalds 已提交
850 851
}

852
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
853 854 855 856 857
void add_disk_randomness(struct gendisk *disk)
{
	if (!disk || !disk->random)
		return;
	/* first major is 1, so we get >= 0x200 here */
858 859
	DEBUG_ENT("disk event %d:%d\n",
		  MAJOR(disk_devt(disk)), MINOR(disk_devt(disk)));
L
Linus Torvalds 已提交
860

861
	add_timer_randomness(disk->random, 0x100 + disk_devt(disk));
L
Linus Torvalds 已提交
862
}
863
#endif
L
Linus Torvalds 已提交
864 865 866 867 868 869 870

/*********************************************************************
 *
 * Entropy extraction routines
 *
 *********************************************************************/

871
static ssize_t extract_entropy(struct entropy_store *r, void *buf,
L
Linus Torvalds 已提交
872 873 874
			       size_t nbytes, int min, int rsvd);

/*
L
Lucas De Marchi 已提交
875
 * This utility inline function is responsible for transferring entropy
L
Linus Torvalds 已提交
876 877 878 879 880
 * from the primary pool to the secondary extraction pool. We make
 * sure we pull enough for a 'catastrophic reseed'.
 */
static void xfer_secondary_pool(struct entropy_store *r, size_t nbytes)
{
881
	__u32	tmp[OUTPUT_POOL_WORDS];
L
Linus Torvalds 已提交
882

883 884 885
	if (r->pull &&
	    r->entropy_count < (nbytes << (ENTROPY_SHIFT + 3)) &&
	    r->entropy_count < r->poolinfo->poolfracbits) {
886
		/* If we're limited, always leave two wakeup worth's BITS */
L
Linus Torvalds 已提交
887
		int rsvd = r->limit ? 0 : random_read_wakeup_thresh/4;
888 889 890 891 892
		int bytes = nbytes;

		/* pull at least as many as BYTES as wakeup BITS */
		bytes = max_t(int, bytes, random_read_wakeup_thresh / 8);
		/* but never more than the buffer size */
893
		bytes = min_t(int, bytes, sizeof(tmp));
L
Linus Torvalds 已提交
894 895

		DEBUG_ENT("going to reseed %s with %d bits "
J
Jiri Kosina 已提交
896
			  "(%zu of %d requested)\n",
897 898
			  r->name, bytes * 8, nbytes * 8,
			  r->entropy_count >> ENTROPY_SHIFT);
L
Linus Torvalds 已提交
899

900
		bytes = extract_entropy(r->pull, tmp, bytes,
901
					random_read_wakeup_thresh / 8, rsvd);
902
		mix_pool_bytes(r, tmp, bytes, NULL);
903
		credit_entropy_bits(r, bytes*8);
L
Linus Torvalds 已提交
904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922
	}
}

/*
 * These functions extracts randomness from the "entropy pool", and
 * returns it in a buffer.
 *
 * The min parameter specifies the minimum amount we can pull before
 * failing to avoid races that defeat catastrophic reseeding while the
 * reserved parameter indicates how much entropy we must leave in the
 * pool after each pull to avoid starving other readers.
 *
 * Note: extract_entropy() assumes that .poolwords is a multiple of 16 words.
 */

static size_t account(struct entropy_store *r, size_t nbytes, int min,
		      int reserved)
{
	unsigned long flags;
923
	int wakeup_write = 0;
924 925 926
	int have_bytes;
	int entropy_count, orig;
	size_t ibytes;
L
Linus Torvalds 已提交
927 928 929 930

	/* Hold lock while accounting */
	spin_lock_irqsave(&r->lock, flags);

931
	BUG_ON(r->entropy_count > r->poolinfo->poolfracbits);
J
Jiri Kosina 已提交
932
	DEBUG_ENT("trying to extract %zu bits from %s\n",
L
Linus Torvalds 已提交
933 934 935
		  nbytes * 8, r->name);

	/* Can we pull enough? */
936
retry:
937 938 939 940 941 942
	entropy_count = orig = ACCESS_ONCE(r->entropy_count);
	have_bytes = entropy_count >> (ENTROPY_SHIFT + 3);
	ibytes = nbytes;
	if (have_bytes < min + reserved) {
		ibytes = 0;
	} else {
L
Linus Torvalds 已提交
943
		/* If limited, never pull more than available */
944 945 946 947 948 949 950
		if (r->limit && ibytes + reserved >= have_bytes)
			ibytes = have_bytes - reserved;

		if (have_bytes >= ibytes + reserved)
			entropy_count -= ibytes << (ENTROPY_SHIFT + 3);
		else
			entropy_count = reserved << (ENTROPY_SHIFT + 3);
951

952 953 954 955 956
		if (cmpxchg(&r->entropy_count, orig, entropy_count) != orig)
			goto retry;

		if ((r->entropy_count >> ENTROPY_SHIFT)
		    < random_write_wakeup_thresh)
957
			wakeup_write = 1;
L
Linus Torvalds 已提交
958 959
	}

J
Jiri Kosina 已提交
960
	DEBUG_ENT("debiting %zu entropy credits from %s%s\n",
961
		  ibytes * 8, r->name, r->limit ? "" : " (unlimited)");
L
Linus Torvalds 已提交
962 963 964

	spin_unlock_irqrestore(&r->lock, flags);

965 966 967 968 969
	if (wakeup_write) {
		wake_up_interruptible(&random_write_wait);
		kill_fasync(&fasync, SIGIO, POLL_OUT);
	}

970
	return ibytes;
L
Linus Torvalds 已提交
971 972 973 974
}

static void extract_buf(struct entropy_store *r, __u8 *out)
{
975
	int i;
976 977
	union {
		__u32 w[5];
978
		unsigned long l[LONGS(20)];
979 980
	} hash;
	__u32 workspace[SHA_WORKSPACE_WORDS];
981
	__u8 extract[64];
982
	unsigned long flags;
L
Linus Torvalds 已提交
983

984
	/* Generate a hash across the pool, 16 words (512 bits) at a time */
985
	sha_init(hash.w);
986
	spin_lock_irqsave(&r->lock, flags);
987
	for (i = 0; i < r->poolinfo->poolwords; i += 16)
988
		sha_transform(hash.w, (__u8 *)(r->pool + i), workspace);
989

990 991 992 993 994 995 996 997 998 999 1000
	/*
	 * If we have a architectural hardware random number
	 * generator, mix that in, too.
	 */
	for (i = 0; i < LONGS(20); i++) {
		unsigned long v;
		if (!arch_get_random_long(&v))
			break;
		hash.l[i] ^= v;
	}

L
Linus Torvalds 已提交
1001
	/*
1002 1003 1004 1005 1006 1007 1008
	 * We mix the hash back into the pool to prevent backtracking
	 * attacks (where the attacker knows the state of the pool
	 * plus the current outputs, and attempts to find previous
	 * ouputs), unless the hash function can be inverted. By
	 * mixing at least a SHA1 worth of hash data back, we make
	 * brute-forcing the feedback as hard as brute-forcing the
	 * hash.
L
Linus Torvalds 已提交
1009
	 */
1010
	__mix_pool_bytes(r, hash.w, sizeof(hash.w), extract);
1011
	spin_unlock_irqrestore(&r->lock, flags);
L
Linus Torvalds 已提交
1012 1013

	/*
1014 1015
	 * To avoid duplicates, we atomically extract a portion of the
	 * pool while mixing, and hash one final time.
L
Linus Torvalds 已提交
1016
	 */
1017
	sha_transform(hash.w, extract, workspace);
1018 1019
	memset(extract, 0, sizeof(extract));
	memset(workspace, 0, sizeof(workspace));
L
Linus Torvalds 已提交
1020 1021

	/*
1022 1023 1024
	 * In case the hash function has some recognizable output
	 * pattern, we fold it in half. Thus, we always feed back
	 * twice as much data as we output.
L
Linus Torvalds 已提交
1025
	 */
1026 1027 1028 1029 1030 1031
	hash.w[0] ^= hash.w[3];
	hash.w[1] ^= hash.w[4];
	hash.w[2] ^= rol32(hash.w[2], 16);

	memcpy(out, &hash, EXTRACT_SIZE);
	memset(&hash, 0, sizeof(hash));
L
Linus Torvalds 已提交
1032 1033
}

1034
static ssize_t extract_entropy(struct entropy_store *r, void *buf,
1035
				 size_t nbytes, int min, int reserved)
L
Linus Torvalds 已提交
1036 1037 1038
{
	ssize_t ret = 0, i;
	__u8 tmp[EXTRACT_SIZE];
1039
	unsigned long flags;
L
Linus Torvalds 已提交
1040

1041
	/* if last_data isn't primed, we need EXTRACT_SIZE extra bytes */
1042 1043 1044 1045 1046 1047
	if (fips_enabled) {
		spin_lock_irqsave(&r->lock, flags);
		if (!r->last_data_init) {
			r->last_data_init = true;
			spin_unlock_irqrestore(&r->lock, flags);
			trace_extract_entropy(r->name, EXTRACT_SIZE,
1048
					      ENTROPY_BITS(r), _RET_IP_);
1049 1050 1051 1052 1053 1054 1055
			xfer_secondary_pool(r, EXTRACT_SIZE);
			extract_buf(r, tmp);
			spin_lock_irqsave(&r->lock, flags);
			memcpy(r->last_data, tmp, EXTRACT_SIZE);
		}
		spin_unlock_irqrestore(&r->lock, flags);
	}
1056

1057
	trace_extract_entropy(r->name, nbytes, ENTROPY_BITS(r), _RET_IP_);
L
Linus Torvalds 已提交
1058 1059 1060 1061 1062
	xfer_secondary_pool(r, nbytes);
	nbytes = account(r, nbytes, min, reserved);

	while (nbytes) {
		extract_buf(r, tmp);
1063

M
Matt Mackall 已提交
1064
		if (fips_enabled) {
1065 1066 1067 1068 1069 1070
			spin_lock_irqsave(&r->lock, flags);
			if (!memcmp(tmp, r->last_data, EXTRACT_SIZE))
				panic("Hardware RNG duplicated output!\n");
			memcpy(r->last_data, tmp, EXTRACT_SIZE);
			spin_unlock_irqrestore(&r->lock, flags);
		}
L
Linus Torvalds 已提交
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
		i = min_t(int, nbytes, EXTRACT_SIZE);
		memcpy(buf, tmp, i);
		nbytes -= i;
		buf += i;
		ret += i;
	}

	/* Wipe data just returned from memory */
	memset(tmp, 0, sizeof(tmp));

	return ret;
}

static ssize_t extract_entropy_user(struct entropy_store *r, void __user *buf,
				    size_t nbytes)
{
	ssize_t ret = 0, i;
	__u8 tmp[EXTRACT_SIZE];

1090
	trace_extract_entropy_user(r->name, nbytes, ENTROPY_BITS(r), _RET_IP_);
L
Linus Torvalds 已提交
1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
	xfer_secondary_pool(r, nbytes);
	nbytes = account(r, nbytes, 0, 0);

	while (nbytes) {
		if (need_resched()) {
			if (signal_pending(current)) {
				if (ret == 0)
					ret = -ERESTARTSYS;
				break;
			}
			schedule();
		}

		extract_buf(r, tmp);
		i = min_t(int, nbytes, EXTRACT_SIZE);
		if (copy_to_user(buf, tmp, i)) {
			ret = -EFAULT;
			break;
		}

		nbytes -= i;
		buf += i;
		ret += i;
	}

	/* Wipe data just returned from memory */
	memset(tmp, 0, sizeof(tmp));

	return ret;
}

/*
 * This function is the exported kernel interface.  It returns some
1124 1125 1126
 * number of good random numbers, suitable for key generation, seeding
 * TCP sequence numbers, etc.  It does not use the hw random number
 * generator, if available; use get_random_bytes_arch() for that.
L
Linus Torvalds 已提交
1127 1128
 */
void get_random_bytes(void *buf, int nbytes)
1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144
{
	extract_entropy(&nonblocking_pool, buf, nbytes, 0, 0);
}
EXPORT_SYMBOL(get_random_bytes);

/*
 * This function will use the architecture-specific hardware random
 * number generator if it is available.  The arch-specific hw RNG will
 * almost certainly be faster than what we can do in software, but it
 * is impossible to verify that it is implemented securely (as
 * opposed, to, say, the AES encryption of a sequence number using a
 * key known by the NSA).  So it's useful if we need the speed, but
 * only if we're willing to trust the hardware manufacturer not to
 * have put in a back door.
 */
void get_random_bytes_arch(void *buf, int nbytes)
L
Linus Torvalds 已提交
1145
{
1146 1147
	char *p = buf;

1148
	trace_get_random_bytes(nbytes, _RET_IP_);
1149 1150 1151
	while (nbytes) {
		unsigned long v;
		int chunk = min(nbytes, (int)sizeof(unsigned long));
1152

1153 1154 1155
		if (!arch_get_random_long(&v))
			break;
		
L
Luck, Tony 已提交
1156
		memcpy(p, &v, chunk);
1157 1158 1159 1160
		p += chunk;
		nbytes -= chunk;
	}

1161 1162
	if (nbytes)
		extract_entropy(&nonblocking_pool, p, nbytes, 0, 0);
L
Linus Torvalds 已提交
1163
}
1164 1165
EXPORT_SYMBOL(get_random_bytes_arch);

L
Linus Torvalds 已提交
1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177

/*
 * init_std_data - initialize pool with system data
 *
 * @r: pool to initialize
 *
 * This function clears the pool's entropy count and mixes some system
 * data into the pool to prepare it for use. The pool is not cleared
 * as that can only decrease the entropy in the pool.
 */
static void init_std_data(struct entropy_store *r)
{
1178
	int i;
1179 1180
	ktime_t now = ktime_get_real();
	unsigned long rv;
L
Linus Torvalds 已提交
1181 1182

	r->entropy_count = 0;
1183
	r->entropy_total = 0;
1184
	r->last_data_init = false;
1185
	mix_pool_bytes(r, &now, sizeof(now), NULL);
1186
	for (i = r->poolinfo->poolbytes; i > 0; i -= sizeof(rv)) {
1187
		if (!arch_get_random_long(&rv))
1188
			break;
1189
		mix_pool_bytes(r, &rv, sizeof(rv), NULL);
1190
	}
1191
	mix_pool_bytes(r, utsname(), sizeof(*(utsname())), NULL);
L
Linus Torvalds 已提交
1192 1193
}

1194 1195 1196 1197 1198 1199 1200 1201 1202 1203
/*
 * Note that setup_arch() may call add_device_randomness()
 * long before we get here. This allows seeding of the pools
 * with some platform dependent data very early in the boot
 * process. But it limits our options here. We must use
 * statically allocated structures that already have all
 * initializations complete at compile time. We should also
 * take care not to overwrite the precious per platform data
 * we were given.
 */
M
Matt Mackall 已提交
1204
static int rand_initialize(void)
L
Linus Torvalds 已提交
1205 1206 1207 1208 1209 1210 1211 1212
{
	init_std_data(&input_pool);
	init_std_data(&blocking_pool);
	init_std_data(&nonblocking_pool);
	return 0;
}
module_init(rand_initialize);

1213
#ifdef CONFIG_BLOCK
L
Linus Torvalds 已提交
1214 1215 1216 1217 1218
void rand_initialize_disk(struct gendisk *disk)
{
	struct timer_rand_state *state;

	/*
1219
	 * If kzalloc returns null, we just won't use that entropy
L
Linus Torvalds 已提交
1220 1221
	 * source.
	 */
1222 1223
	state = kzalloc(sizeof(struct timer_rand_state), GFP_KERNEL);
	if (state)
L
Linus Torvalds 已提交
1224 1225
		disk->random = state;
}
1226
#endif
L
Linus Torvalds 已提交
1227 1228

static ssize_t
1229
random_read(struct file *file, char __user *buf, size_t nbytes, loff_t *ppos)
L
Linus Torvalds 已提交
1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
{
	ssize_t n, retval = 0, count = 0;

	if (nbytes == 0)
		return 0;

	while (nbytes > 0) {
		n = nbytes;
		if (n > SEC_XFER_SIZE)
			n = SEC_XFER_SIZE;

J
Jiri Kosina 已提交
1241
		DEBUG_ENT("reading %zu bits\n", n*8);
L
Linus Torvalds 已提交
1242 1243 1244

		n = extract_entropy_user(&blocking_pool, buf, n);

J
Jiri Kosina 已提交
1245 1246 1247 1248 1249 1250
		if (n < 0) {
			retval = n;
			break;
		}

		DEBUG_ENT("read got %zd bits (%zd still needed)\n",
L
Linus Torvalds 已提交
1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261
			  n*8, (nbytes-n)*8);

		if (n == 0) {
			if (file->f_flags & O_NONBLOCK) {
				retval = -EAGAIN;
				break;
			}

			DEBUG_ENT("sleeping?\n");

			wait_event_interruptible(random_read_wait,
1262 1263
				ENTROPY_BITS(&input_pool) >=
				random_read_wakeup_thresh);
L
Linus Torvalds 已提交
1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285

			DEBUG_ENT("awake\n");

			if (signal_pending(current)) {
				retval = -ERESTARTSYS;
				break;
			}

			continue;
		}

		count += n;
		buf += n;
		nbytes -= n;
		break;		/* This break makes the device work */
				/* like a named pipe */
	}

	return (count ? count : retval);
}

static ssize_t
1286
urandom_read(struct file *file, char __user *buf, size_t nbytes, loff_t *ppos)
L
Linus Torvalds 已提交
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
{
	return extract_entropy_user(&nonblocking_pool, buf, nbytes);
}

static unsigned int
random_poll(struct file *file, poll_table * wait)
{
	unsigned int mask;

	poll_wait(file, &random_read_wait, wait);
	poll_wait(file, &random_write_wait, wait);
	mask = 0;
1299
	if (ENTROPY_BITS(&input_pool) >= random_read_wakeup_thresh)
L
Linus Torvalds 已提交
1300
		mask |= POLLIN | POLLRDNORM;
1301
	if (ENTROPY_BITS(&input_pool) < random_write_wakeup_thresh)
L
Linus Torvalds 已提交
1302 1303 1304 1305
		mask |= POLLOUT | POLLWRNORM;
	return mask;
}

1306 1307
static int
write_pool(struct entropy_store *r, const char __user *buffer, size_t count)
L
Linus Torvalds 已提交
1308 1309 1310 1311 1312
{
	size_t bytes;
	__u32 buf[16];
	const char __user *p = buffer;

1313 1314 1315 1316
	while (count > 0) {
		bytes = min(count, sizeof(buf));
		if (copy_from_user(&buf, p, bytes))
			return -EFAULT;
L
Linus Torvalds 已提交
1317

1318
		count -= bytes;
L
Linus Torvalds 已提交
1319 1320
		p += bytes;

1321
		mix_pool_bytes(r, buf, bytes, NULL);
1322
		cond_resched();
L
Linus Torvalds 已提交
1323
	}
1324 1325 1326 1327

	return 0;
}

1328 1329
static ssize_t random_write(struct file *file, const char __user *buffer,
			    size_t count, loff_t *ppos)
1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
{
	size_t ret;

	ret = write_pool(&blocking_pool, buffer, count);
	if (ret)
		return ret;
	ret = write_pool(&nonblocking_pool, buffer, count);
	if (ret)
		return ret;

	return (ssize_t)count;
L
Linus Torvalds 已提交
1341 1342
}

M
Matt Mackall 已提交
1343
static long random_ioctl(struct file *f, unsigned int cmd, unsigned long arg)
L
Linus Torvalds 已提交
1344 1345 1346 1347 1348 1349 1350
{
	int size, ent_count;
	int __user *p = (int __user *)arg;
	int retval;

	switch (cmd) {
	case RNDGETENTCNT:
M
Matt Mackall 已提交
1351
		/* inherently racy, no point locking */
1352 1353
		ent_count = ENTROPY_BITS(&input_pool);
		if (put_user(ent_count, p))
L
Linus Torvalds 已提交
1354 1355 1356 1357 1358 1359 1360
			return -EFAULT;
		return 0;
	case RNDADDTOENTCNT:
		if (!capable(CAP_SYS_ADMIN))
			return -EPERM;
		if (get_user(ent_count, p))
			return -EFAULT;
1361
		credit_entropy_bits_safe(&input_pool, ent_count);
L
Linus Torvalds 已提交
1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
		return 0;
	case RNDADDENTROPY:
		if (!capable(CAP_SYS_ADMIN))
			return -EPERM;
		if (get_user(ent_count, p++))
			return -EFAULT;
		if (ent_count < 0)
			return -EINVAL;
		if (get_user(size, p++))
			return -EFAULT;
1372 1373
		retval = write_pool(&input_pool, (const char __user *)p,
				    size);
L
Linus Torvalds 已提交
1374 1375
		if (retval < 0)
			return retval;
1376
		credit_entropy_bits_safe(&input_pool, ent_count);
L
Linus Torvalds 已提交
1377 1378 1379 1380 1381 1382
		return 0;
	case RNDZAPENTCNT:
	case RNDCLEARPOOL:
		/* Clear the entropy pool counters. */
		if (!capable(CAP_SYS_ADMIN))
			return -EPERM;
M
Matt Mackall 已提交
1383
		rand_initialize();
L
Linus Torvalds 已提交
1384 1385 1386 1387 1388 1389
		return 0;
	default:
		return -EINVAL;
	}
}

1390 1391 1392 1393 1394
static int random_fasync(int fd, struct file *filp, int on)
{
	return fasync_helper(fd, filp, on, &fasync);
}

1395
const struct file_operations random_fops = {
L
Linus Torvalds 已提交
1396 1397 1398
	.read  = random_read,
	.write = random_write,
	.poll  = random_poll,
M
Matt Mackall 已提交
1399
	.unlocked_ioctl = random_ioctl,
1400
	.fasync = random_fasync,
1401
	.llseek = noop_llseek,
L
Linus Torvalds 已提交
1402 1403
};

1404
const struct file_operations urandom_fops = {
L
Linus Torvalds 已提交
1405 1406
	.read  = urandom_read,
	.write = random_write,
M
Matt Mackall 已提交
1407
	.unlocked_ioctl = random_ioctl,
1408
	.fasync = random_fasync,
1409
	.llseek = noop_llseek,
L
Linus Torvalds 已提交
1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
};

/***************************************************************
 * Random UUID interface
 *
 * Used here for a Boot ID, but can be useful for other kernel
 * drivers.
 ***************************************************************/

/*
 * Generate random UUID
 */
void generate_random_uuid(unsigned char uuid_out[16])
{
	get_random_bytes(uuid_out, 16);
1425
	/* Set UUID version to 4 --- truly random generation */
L
Linus Torvalds 已提交
1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455
	uuid_out[6] = (uuid_out[6] & 0x0F) | 0x40;
	/* Set the UUID variant to DCE */
	uuid_out[8] = (uuid_out[8] & 0x3F) | 0x80;
}
EXPORT_SYMBOL(generate_random_uuid);

/********************************************************************
 *
 * Sysctl interface
 *
 ********************************************************************/

#ifdef CONFIG_SYSCTL

#include <linux/sysctl.h>

static int min_read_thresh = 8, min_write_thresh;
static int max_read_thresh = INPUT_POOL_WORDS * 32;
static int max_write_thresh = INPUT_POOL_WORDS * 32;
static char sysctl_bootid[16];

/*
 * These functions is used to return both the bootid UUID, and random
 * UUID.  The difference is in whether table->data is NULL; if it is,
 * then a new UUID is generated and returned to the user.
 *
 * If the user accesses this via the proc interface, it will be returned
 * as an ASCII string in the standard UUID format.  If accesses via the
 * sysctl system call, it is returned as 16 bytes of binary data.
 */
1456
static int proc_do_uuid(struct ctl_table *table, int write,
L
Linus Torvalds 已提交
1457 1458
			void __user *buffer, size_t *lenp, loff_t *ppos)
{
1459
	struct ctl_table fake_table;
L
Linus Torvalds 已提交
1460 1461 1462 1463 1464 1465
	unsigned char buf[64], tmp_uuid[16], *uuid;

	uuid = table->data;
	if (!uuid) {
		uuid = tmp_uuid;
		generate_random_uuid(uuid);
1466 1467 1468 1469 1470 1471 1472 1473
	} else {
		static DEFINE_SPINLOCK(bootid_spinlock);

		spin_lock(&bootid_spinlock);
		if (!uuid[8])
			generate_random_uuid(uuid);
		spin_unlock(&bootid_spinlock);
	}
L
Linus Torvalds 已提交
1474

J
Joe Perches 已提交
1475 1476
	sprintf(buf, "%pU", uuid);

L
Linus Torvalds 已提交
1477 1478 1479
	fake_table.data = buf;
	fake_table.maxlen = sizeof(buf);

1480
	return proc_dostring(&fake_table, write, buffer, lenp, ppos);
L
Linus Torvalds 已提交
1481 1482
}

1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499
/*
 * Return entropy available scaled to integral bits
 */
static int proc_do_entropy(ctl_table *table, int write,
			   void __user *buffer, size_t *lenp, loff_t *ppos)
{
	ctl_table fake_table;
	int entropy_count;

	entropy_count = *(int *)table->data >> ENTROPY_SHIFT;

	fake_table.data = &entropy_count;
	fake_table.maxlen = sizeof(entropy_count);

	return proc_dointvec(&fake_table, write, buffer, lenp, ppos);
}

L
Linus Torvalds 已提交
1500
static int sysctl_poolsize = INPUT_POOL_WORDS * 32;
1501 1502
extern struct ctl_table random_table[];
struct ctl_table random_table[] = {
L
Linus Torvalds 已提交
1503 1504 1505 1506 1507
	{
		.procname	= "poolsize",
		.data		= &sysctl_poolsize,
		.maxlen		= sizeof(int),
		.mode		= 0444,
1508
		.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
1509 1510 1511 1512 1513
	},
	{
		.procname	= "entropy_avail",
		.maxlen		= sizeof(int),
		.mode		= 0444,
1514
		.proc_handler	= proc_do_entropy,
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520 1521
		.data		= &input_pool.entropy_count,
	},
	{
		.procname	= "read_wakeup_threshold",
		.data		= &random_read_wakeup_thresh,
		.maxlen		= sizeof(int),
		.mode		= 0644,
1522
		.proc_handler	= proc_dointvec_minmax,
L
Linus Torvalds 已提交
1523 1524 1525 1526 1527 1528 1529 1530
		.extra1		= &min_read_thresh,
		.extra2		= &max_read_thresh,
	},
	{
		.procname	= "write_wakeup_threshold",
		.data		= &random_write_wakeup_thresh,
		.maxlen		= sizeof(int),
		.mode		= 0644,
1531
		.proc_handler	= proc_dointvec_minmax,
L
Linus Torvalds 已提交
1532 1533 1534 1535 1536 1537 1538 1539
		.extra1		= &min_write_thresh,
		.extra2		= &max_write_thresh,
	},
	{
		.procname	= "boot_id",
		.data		= &sysctl_bootid,
		.maxlen		= 16,
		.mode		= 0444,
1540
		.proc_handler	= proc_do_uuid,
L
Linus Torvalds 已提交
1541 1542 1543 1544 1545
	},
	{
		.procname	= "uuid",
		.maxlen		= 16,
		.mode		= 0444,
1546
		.proc_handler	= proc_do_uuid,
L
Linus Torvalds 已提交
1547
	},
1548
	{ }
L
Linus Torvalds 已提交
1549 1550 1551
};
#endif 	/* CONFIG_SYSCTL */

1552
static u32 random_int_secret[MD5_MESSAGE_BYTES / 4] ____cacheline_aligned;
L
Linus Torvalds 已提交
1553

1554
int random_int_secret_init(void)
L
Linus Torvalds 已提交
1555
{
1556
	get_random_bytes(random_int_secret, sizeof(random_int_secret));
L
Linus Torvalds 已提交
1557 1558 1559 1560 1561 1562 1563 1564 1565
	return 0;
}

/*
 * Get a random word for internal kernel use only. Similar to urandom but
 * with the goal of minimal entropy pool depletion. As a result, the random
 * value is not cryptographically secure but for several uses the cost of
 * depleting entropy is too high
 */
T
Theodore Ts'o 已提交
1566
static DEFINE_PER_CPU(__u32 [MD5_DIGEST_WORDS], get_random_int_hash);
L
Linus Torvalds 已提交
1567 1568
unsigned int get_random_int(void)
{
1569
	__u32 *hash;
1570
	unsigned int ret;
1571

1572 1573 1574 1575
	if (arch_get_random_int(&ret))
		return ret;

	hash = get_cpu_var(get_random_int_hash);
1576

1577
	hash[0] += current->pid + jiffies + random_get_entropy();
1578 1579
	md5_transform(hash, random_int_secret);
	ret = hash[0];
1580 1581 1582
	put_cpu_var(get_random_int_hash);

	return ret;
L
Linus Torvalds 已提交
1583
}
1584
EXPORT_SYMBOL(get_random_int);
L
Linus Torvalds 已提交
1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603

/*
 * randomize_range() returns a start address such that
 *
 *    [...... <range> .....]
 *  start                  end
 *
 * a <range> with size "len" starting at the return value is inside in the
 * area defined by [start, end], but is otherwise randomized.
 */
unsigned long
randomize_range(unsigned long start, unsigned long end, unsigned long len)
{
	unsigned long range = end - len - start;

	if (end <= start + len)
		return 0;
	return PAGE_ALIGN(get_random_int() % range + start);
}